JMIR Publications

ABSTRACT

Background:

Firearm violence injury is captured via structured data codes that best reflect acute bodily injury. There are no structured data codes to describe secondary exposure (e.g. witnessing a shooting, being threatened by a firearm, or losing a loved one to gun violence and injury from firearms) even though such exposure is associated with many long and short term health impacts. Clinical chart notes from Electronic Health Records (EHRs) often contain data not otherwise captured in structured data fields and can be categorized using natural language processing (NLP).

Objective:

The study protocol described here outlines the steps being taken to develop an NLP text classifier for determination of exposure to firearm violence from ambulatory primary care and behavioral health EHR clinical progress notes for persons aged ≥ 5 years.

Methods:

describe the process for arriving at a novel NLP lexicon, clinical process note selection, the steps for text classifier training and selection, and evaluation of model performance. We also describe involvement of a stakeholder advisory committee in the development of the lexicon, and how the lexicon addresses biases inherent in NLP text classifiers.

Results:

We describe the development of an NLP lexicon with the input of a stakeholder advisory committee, evaluation of the text classifier, and future plans for utilization of the text classifier.

Conclusions:

This work describes the development of a novel NLP text classifier to identify exposure to firearm violence in ambulatory primary care and behavioral health clinical progress notes. Clinical Trial: This is an IRB exempt non-intervention data study and was not registered on clinicaltrials.gov